Early burnout definitions emphasized worker exhaustion resulting from accumulated life stress (Guthier et al., 2020). Contemporary perspectives frame burnout as a job-induced human factor shaped by personal experiences and existential work engagement (Adebayo, 2022; Jooste, 2020). Public training increasingly targets symptom recognition and motivation, while burnout inventories align measurement with these evolving conceptualizations (Shoman et al., 2021; Schaufeli et al., 2020).

This study draws on established theoretical models explaining the nature and origins of burnout (Galanakis & Tsitouri, 2022; Prentice et al., 2023). While the Maslach Burnout Inventory remains the classical measure (Aguayo-Estremera et al., 2023), newer frameworks, such as the Job Demands–Resources and Person–Environment Fit models, reflect conditions in developed societies (Begum et al., 2024; Xiong et al., 2022). These models posit that burnout arises from excessive demands and insufficient resources (Edú-Valsania et al., 2022; Bakker & de Vries, 2021; Bakker et al., 2023).

Each selected model provides distinct indicators for understanding the persistence and variation of burnout in professional settings, though each has strengths and limitations. These models strongly reflect the individual, social, and economic realities of Sudanese schools (Macaulay, 2023; Hamid et al., 2020). While applied primarily to education, their relevance extends across sectors, offering a foundation for identifying measurable indicators and informing prevention-oriented policy strategies (Oke & Fernandes, 2020).

The development of CBI was not an immediate process but emerged from extensive prior research on burnout. While the Maslach Burnout Inventory (MBI) originated in the mid-1970s, major theoretical and empirical advances across diverse fields, including healthcare, education, management, sports, air-traffic control, and human service work, accelerated during the 1980s and 1990s (Wilson, 2020; Vila et al., 2022). Numerous studies demonstrated significant associations between MBI scores and multiple variables, emphasizing the need for a refined and valid alternative measure (Obregon et al., 2020; González-Rodríguez et al., 2020; Soares et al., 2023).

Through extensive qualitative research, including in-depth interviews, participant observation, and field testing across Danish occupational sectors, the CBI was systematically developed and finalized (Thrush et al., 2021; Fadare et al., 2021; Montgomery et al., 2021; Piperac et al., 2021). The instrument was designed to assess three core dimensions identified through intervention studies: personal burnout, work-related burnout, and client-related burnout (Fadare et al., 2021; Barton et al., 2022; Alvey, 2020; Tran et al., 2023). These dimensions reflect a cross-cultural, subjective understanding of burnout and aim to capture its full spectrum of consequences.

Pilot testing of eight CBI versions in Sudan demonstrated partial factorial validity, item clarity, and strong intercorrelations among the three scales, supporting its relevance in varied occupational settings. The CBI thus provides health professionals and organizational leaders with a practical tool for identifying, diagnosing, and addressing burnout across professions (Montgomery et al., 2021; Majeed et al., 2022; Piperac et al., 2021).

The primary aim of this study is to examine the psychometric properties of CBI in assessing burnout among teachers and other workers in Sudan. By evaluating its validity and reliability within Sudan’s cultural, social, and professional context, the study seeks to provide insights into burnout experiences across demanding occupations and to strengthen the understanding of occupational stress in the region, benefiting both educational and workplace sectors.

The study was conducted in the 2021–2022 academic year in Khartoum, the capital of Sudan, in its 7 localities: Al-Khartoum, Jabal-olia, Khartoum-bahri, Shariq-Alniel, Omdurman, Umbada, and Karrari. the number of participants were (1068), males (n = 461; 43.2%), females (n = 607; 56.8%); they were distributed in three different professions: Bank employees (n = 304; 28.5%), medical staff (n = 326; 30.5%), and teachers (n = 438; 41%). Their qualifications varied: less than Secondary School Certificate (n = 6; 0.6%), Secondary School Certificate (n = 43; 4%), Middle Diploma (n = 70; 6.5%), Bachelor (n = 698, 65.4%), Higher Diploma (n = 80; 7.5%), Master Degree (n = 138; 12.9%), Doctorate holders (n = 33; 3.1%).

Data were collected using the CBI, distributed by institutional volunteers. Ethical approval was granted by the University of Khartoum, Faculty of Arts Research Board (No. 4–81, Oct 2022). Participants provided informed consent and were assured confidentiality, with all procedures conducted in accordance with relevant guidelines and regulations.

The original CBI was obtained from Kristensen et al. (2005), and permission was granted by the authors to adapt it culturally and linguistically for use in Sudan. The scale was first translated from English to Arabic by two university professors specializing in psychology and English language teaching. The authors then revised the translation. A back-translation into English was conducted by another professor specializing in English–Arabic linguistics, confirming equivalence between versions. The Arabic version was subsequently reviewed by 11 experts in mental health and organizational psychology, who recommended simplifying the language for everyday use and converting items from questions to statements to align with a five-point Likert scale. All suggested revisions were incorporated, establishing surface validity for the adapted instrument.

A pilot test involving 103 participants was conducted to evaluate the expert panel’s recommendations. Reliability for the three CBI factors was acceptable (α = 0.701, 0.745, and 0.723). Validity, assessed using the square root of alpha, yielded √α values of 0.837, 0.863, and 0.850, indicating satisfactory psychometric properties (AL-Qadri & Zhao, 2021; Smits et al., 2018). The final data collection used 19 items across three domains: PB (6 items), WB (7 items), and CB (6 items), with sample items such as “I experience physical exhaustion during work” and “Working with students drains my energy.”

A team of 20 volunteers with a bachelor’s degree in psychology was formed, and the following steps were undertaken: First, the data collection team was trained. Second, the necessary approvals were obtained prior to data collection. Third, the volunteers were allocated to each occupational group. Fourth, the questionnaires were administered, checked for completeness, scored, and entered an electronic database over a four-month period to complete the cross-sectional process. Finally, the findings were compiled and reported in the manuscript.

A principal components analysis with varimax rotation was conducted for teachers, medical staff, and bank employees. Kaiser–Meyer–Olkin (KMO) values (0.926, 0.930, 0.947) exceeded the 0.60 threshold (Kaiser, 1974; Field, 2009). Bartlett’s tests were significant (χ² = 4299.793, 4071.961, 6446.505, p < 0.001), confirming factorability as shown in Appendix A.

All items showed communalities above 0.40 across the three factors for teachers, medical staff, and bank employees. Three factors emerged for each group, explaining 58.146%, 63.820%, and 80.285% of variance, respectively. Items were assigned to factors based on loadings >0.40; cross-loading items were allocated to the factor with the highest loading (González-Rodríguez et al., 2020). All 19 items demonstrated significant loadings (See Appendix A). For teachers, factor loadings ranged from 0.984–0.535 (PB), 0.995–0.751 (WB), and 0.952–0.940 (CB). For medical staff, loadings ranged from 0.899–0.762, 0.862–0.690, and 0.819–0.444, respectively. For bank employees, loadings ranged from 0.778–0.419, 0.718–0.401, and 0.812–0.503, respectively. Descriptive statistics for all items are reported in Appendix A.

Next, to confirm the orientation of the items and verify the validation for the three factors, researchers used the distribution obtained through Table 1 to show that the reliability coefficients (ω, α, and CR) were all above 0.70, and the AVE values exceeded 0.50, indicating acceptable reliability and discriminant validity according to Tran et al. (2023) and AL-Qadri & Zhao (2021). As well as the results reported, the squared correlations between latent constructs were assessed to assess discriminant validity more rigorously. AVE values for each construct exceeded their corresponding shared variances, confirming both convergent and discriminant validity.

The CBI Scale was refined through confirmatory factor analysis (CFA), focusing on items with high modification indices. All items were retained, as factor loadings exceeded 0.40 (Figure 1a). Initial model fit was inadequate (χ²/df = 5.039; CFI = 0.856; GFI = 0.812; TLI = 0.836; RMSEA = 0.096). Modification indices recommended correlating selected error terms, items 1–2, 8–9, 9–11, 16–18, and 15–19 (Figure 1b). After these modifications, the teachers’ model demonstrated satisfactory fit (χ²/df = 2.854; CFI = 0.936; GFI = 0.921; TLI = 0.925; RMSEA = 0.065), with all loadings remaining above 0.40, confirming model adequacy (Figure 1b).

Likewise, CFA validated the medical staff responses, confirming construct reliability, AVE, and discriminant validity (Table 2). Reliability coefficients (ω, α, CR) exceeded 0.70, and AVE values were above 0.50. Squared inter-factor correlations were lower than each construct’s AVE, demonstrating acceptable convergent and discriminant validity.

The initial CFA model for the medical staff group showed adequate item loadings (Figure 2a) but poor overall fit (χ²/df = 3.624, CFI = 0.912, GFI = 0.857, TLI = 0.900, RMSEA = 0.090), as later detailed in Table 3. Modification indices recommended removing Items 9 and 13 from WB and correlating selected error terms (Items 15–16 and 18–19) to improve fit (Figure 2b). The revised model demonstrated satisfactory indices (χ²/df = 2.942, CFI = 0.944, GFI = 0.898, TLI = 0.934, RMSEA = 0.077). Although Item 13 had a strong loading (0.88), it impaired overall model fit (CFI and TLI < 0.90). Consistent with Kline (2016) and Hair et al. (2019), priority was given to achieving robust model fit, supporting the decision to remove items despite high loadings.

The bank employees’ model was validated using CFA to assess reliability, AVE, and discriminant validity. As shown in Table 4, reliability coefficients (ω, α, CR) exceeded 0.70, and AVE values were above 0.50, meeting criteria recommended by Thrush et al. (2021) and Al-Khadher et al. (2024). Squared inter-construct correlations were lower than each construct’s AVE, confirming adequate convergent and discriminant validity.

The CBI Scale for bank employees was refined through CFA, focusing on items with high modification indices. All items were initially retained because their loadings exceeded 0.40 (Figure 3a). However, the initial model showed inadequate fit (χ²/df = 3.455; CFI = 0.943; GFI = 0.849; TLI = 0.935; RMSEA = 0.090), as later reported in Table 3. Modification indices indicated the need to correlate selected error terms, including Items 1–2 and 2–6 (Factor 1) and Items 12–13 (Factor 2), as shown in Figure 3b. After these adjustments, the model achieved satisfactory fit (χ²/df = 3.089; CFI = 0.952; GFI = 0.891; TLI = 0.944; RMSEA = 0.080). All item loadings remained above 0.40 in the revised model, confirming model adequacy (Figure 3b).

The overall model, comprising all three groups, was validated using CFA to assess reliability, AVE, and discriminant validity. Table 5 shows that reliability coefficients (ω, α, CR) exceeded 0.70 and AVE values were above 0.50, meeting criteria by Thrush et al. (2021) and Al-Khadher et al. (2024). Squared inter-construct correlations were lower than each construct’s AVE, confirming convergent and discriminant validity.

The final CBI Scale was refined through CFA, focusing on items with high modification indices. All items were initially retained because their loadings exceeded 0.40 as shown in Figure 4a. However, the initial model showed inadequate fit (χ²/df = 6.867; CFI = 0.948; GFI = 0.888; TLI = 0.940; RMSEA = 0.074), as reported in Table 3. Modification indices recommended correlating selected error terms, Items 1–2 (Factor 1) and Items 15–19 (Factor 3), to address residual misspecifications (Figure 4b). After these adjustments, the model achieved satisfactory fit (χ²/df = 4.811; CFI = 0.967; GFI = 0.922; TLI = 0.961; RMSEA = 0.060). All item loadings in the revised model remained above 0.40, confirming model adequacy as shown in Figure 4b.

Besides, the researchers provided a study model for each group and an overall model encompassing 19 items for the CBI, except for the medical staff group, which included 17 items. As shown in Table 3, all models’ fits were acceptable after modifications. The overall model was determined to be appropriate for Sudanese workers across various fields, including education, healthcare, and banking. The study’s results indicate that the latest version of the CBI scale, which incorporates all adaptation processes detailed in the methodology, includes 17 items specifically for workers in healthcare. However, other workers can assess their burnout inventory using the latest version of the CBI scale, which is clarified and modified in this study and contains 19-item.

Concurrent validity was assessed by correlating the CBI with the Maslach Burnout Inventory (MBI), including Emotional Exhaustion (EE), Personal Realization (PR), and Depersonalization (D), defined as a detachment from oneself (American Psychiatric Association, 2022) and marked by emotional numbing and disrupted self-awareness (Sierra & David, 2011), and Interpersonal Psychological Stress (IPS) adapted by González-Rodríguez et al. (2020). A total of 242 teachers, medical staff, and bank employees participated.

The investigation of the CBI subscales revealed statistically significant correlations, ranging from [r = 0.365(CB and PB) to 0.729 (WB and PB)], at the 0.05 and 0.01 levels, as shown in Table 6. Moreover, CBI subscales’ values correlated positively significant, ranging from (r = 0.324 to 434), negatively significant, ranging from [r = −0.318 (PR and PB) to −0.475 (PR and CB)], and weakly [r = 0.006 (D and PB) to 0.189 (IPS and PB)] scores of the MBI subscales. Emotional Exhaustion (EE) correlated positively significant the PB, WB, and CB [r = 0.326, 0.347, and 0.434], respectively. Personal Realization (PR) correlated negatively significant the PB, WB, and CB (r = −0.318, −0.438, and −0.475), respectively. While Depersonalization (D) correlated positively significant the CB (r = 0.331). Furthermore, Interpersonal Psychological Stress (IPS) correlated significantly positive the WB and CB (r = 0.369 and 0.324), respectively. These results indicated that there is a good concurrent validity of CBI scale with MBI scale as presented in Table 6.